Matches in SemOpenAlex for { <https://semopenalex.org/work/W4311993311> ?p ?o ?g. }
Showing items 1 to 71 of
71
with 100 items per page.
- W4311993311 abstract "Modeling noise transition matrix is a kind of promising method for learning with label noise. Based on the estimated noise transition matrix and the noisy posterior probabilities, the clean posterior probabilities, which are jointly called Label Distribution (LD) in this paper, can be calculated as the supervision. To reliably estimate the noise transition matrix, some methods assume that anchor points are available during training. Nonetheless, if anchor points are invalid, the noise transition matrix might be poorly learned, resulting in poor performance. Consequently, other methods treat reliable data points, extracted from training data, as pseudo anchor points. However, from a statistical point of view, the noise transition matrix can be inferred from data with noisy labels under the clean-label-domination assumption. Therefore, we aim to estimate the noise transition matrix without (pseudo) anchor points. There is evidence showing that samples are more likely to be mislabeled as other similar class labels, which means the mislabeling probability is highly correlated with the inter-class correlation. Inspired by this observation, we propose an instance-specific Label Distribution Regularization (LDR), in which the instance-specific LD is estimated as the supervision, to prevent DCNNs from memorizing noisy labels. Specifically, we estimate the noisy posterior under the supervision of noisy labels, and approximate the batch-level noise transition matrix by estimating the inter-class correlation matrix with neither anchor points nor pseudo anchor points. Experimental results on two synthetic noisy datasets and two real-world noisy datasets demonstrate that our LDR outperforms existing methods." @default.
- W4311993311 created "2023-01-03" @default.
- W4311993311 creator A5034520275 @default.
- W4311993311 creator A5045416617 @default.
- W4311993311 creator A5052354174 @default.
- W4311993311 creator A5083195196 @default.
- W4311993311 date "2022-12-16" @default.
- W4311993311 modified "2023-09-30" @default.
- W4311993311 title "Instance-specific Label Distribution Regularization for Learning with Label Noise" @default.
- W4311993311 doi "https://doi.org/10.48550/arxiv.2212.08380" @default.
- W4311993311 hasPublicationYear "2022" @default.
- W4311993311 type Work @default.
- W4311993311 citedByCount "0" @default.
- W4311993311 crossrefType "posted-content" @default.
- W4311993311 hasAuthorship W4311993311A5034520275 @default.
- W4311993311 hasAuthorship W4311993311A5045416617 @default.
- W4311993311 hasAuthorship W4311993311A5052354174 @default.
- W4311993311 hasAuthorship W4311993311A5083195196 @default.
- W4311993311 hasBestOaLocation W43119933111 @default.
- W4311993311 hasConcept C106487976 @default.
- W4311993311 hasConcept C11413529 @default.
- W4311993311 hasConcept C115961682 @default.
- W4311993311 hasConcept C119857082 @default.
- W4311993311 hasConcept C153180895 @default.
- W4311993311 hasConcept C154945302 @default.
- W4311993311 hasConcept C159985019 @default.
- W4311993311 hasConcept C192562407 @default.
- W4311993311 hasConcept C21080849 @default.
- W4311993311 hasConcept C2524010 @default.
- W4311993311 hasConcept C2776135515 @default.
- W4311993311 hasConcept C2777212361 @default.
- W4311993311 hasConcept C28719098 @default.
- W4311993311 hasConcept C33923547 @default.
- W4311993311 hasConcept C41008148 @default.
- W4311993311 hasConcept C49555168 @default.
- W4311993311 hasConcept C98763669 @default.
- W4311993311 hasConcept C99498987 @default.
- W4311993311 hasConceptScore W4311993311C106487976 @default.
- W4311993311 hasConceptScore W4311993311C11413529 @default.
- W4311993311 hasConceptScore W4311993311C115961682 @default.
- W4311993311 hasConceptScore W4311993311C119857082 @default.
- W4311993311 hasConceptScore W4311993311C153180895 @default.
- W4311993311 hasConceptScore W4311993311C154945302 @default.
- W4311993311 hasConceptScore W4311993311C159985019 @default.
- W4311993311 hasConceptScore W4311993311C192562407 @default.
- W4311993311 hasConceptScore W4311993311C21080849 @default.
- W4311993311 hasConceptScore W4311993311C2524010 @default.
- W4311993311 hasConceptScore W4311993311C2776135515 @default.
- W4311993311 hasConceptScore W4311993311C2777212361 @default.
- W4311993311 hasConceptScore W4311993311C28719098 @default.
- W4311993311 hasConceptScore W4311993311C33923547 @default.
- W4311993311 hasConceptScore W4311993311C41008148 @default.
- W4311993311 hasConceptScore W4311993311C49555168 @default.
- W4311993311 hasConceptScore W4311993311C98763669 @default.
- W4311993311 hasConceptScore W4311993311C99498987 @default.
- W4311993311 hasLocation W43119933111 @default.
- W4311993311 hasOpenAccess W4311993311 @default.
- W4311993311 hasPrimaryLocation W43119933111 @default.
- W4311993311 hasRelatedWork W1560139384 @default.
- W4311993311 hasRelatedWork W1978450727 @default.
- W4311993311 hasRelatedWork W2030105136 @default.
- W4311993311 hasRelatedWork W2033914206 @default.
- W4311993311 hasRelatedWork W2146076056 @default.
- W4311993311 hasRelatedWork W2163831990 @default.
- W4311993311 hasRelatedWork W2357111437 @default.
- W4311993311 hasRelatedWork W2378160586 @default.
- W4311993311 hasRelatedWork W3003836766 @default.
- W4311993311 hasRelatedWork W4244943737 @default.
- W4311993311 isParatext "false" @default.
- W4311993311 isRetracted "false" @default.
- W4311993311 workType "article" @default.